PP8 KAM 2015 Serial No A-LVV-O-LVC STATISTICS Paper—IIl Time Allowed : Three Hours See) INSTRUCTIONS Please read each of the following instructions carefully before attempting questions : There are EIGHT questions divided under TWO Sections. Candidate has to attempt FIVE questions in all. Question no. 1 and § are compulsory and ou: of the remaining, THREE are to be attempted choosing at least ONE from each Section. The number of marks carried by a question / part is irdicated against it. Unless other-wise mentioned, symbols and notations have their usual standard meanings. Assume suitable data, if necessary and indicate the same clearly. Candidates should attempt questions/parts as per the instructions given in the Section. All parts and sub-parts of a question are to be attempted together in the answer book. Attempts of questions shall be counted in chronological order. Unless struck off, attempt of a question shall be counted even if attempted partly. Any page or portion of the page left blank in the answer book must be clearly struck off. Answers must be written in ENGLISH only, 1 .SECTION—A 1. Answer any FIVE parts from the fcllowing : 5x8=<¢( (a) Show that for SRSWOR, thz sample proportion is an unbiased estimator of at tT. ieS 1 N of the parametric function > i=l where |,’s are known real numbers, not all zero. 10 4 (Contd.@) (a) (b) (c) For a two-stage sampling design, let n villages be selected with probabilities proportional to a given size and with replacement from a populatior. of N villages. From the it selected village consisting of M; households (hh), m, housel.olds are selected by simple random sampling without replacement. Write down an unbiased estimator nM, of y=) s Y;> where Y; is the value of a i=l jel study variate for the j"" hh of the i” village Also write down an unbiased estimator of Var (Y). 10 Describe how uniformity trials in experimental designs and pilot studies in sample surveys are conducted, mentioning their uses. 10 Find the C-matrix for the following design and obtain independent estimable treatment contrasts : B, = (1, 2, 2), B, = (2, 4, 4), B, =(1, 1,2), B,= G4, 4). 10 Explain what is meant by a split-plot design. Suppose that factor A has p levels which are arranged in a Randomized Block Desigr. having r replicates: Let factor B have q levels which are 5 (Contd.)(4) (a) applied to plots of a block after subdividing each plot into q sub plots. Write down the model, cleatly explaining the notations and assumptions anc present a blank ANOVA table with sources of variation and degrees of freedom. 10 State the condi-ions to be satisfied by a Partially Balanced Incomplete Block (PBIB) design. Prove the following restrictions of a PBIBD : m i) Di ajayek jel a (id) 1K -D= Bi neh; 16 jel What are the advantages of stratified sampling ? Consider the allccation of sample size n, to strata given by nx N,9° > Where N, is the stratum size. G, is the within s:ratum standard deviation of the ith stratum, (i = 1, 2, ...., k) and é is real. Write down an expression for the Var (Y. “.) Where Yo st. stands for the unb_ased estimator of the population mean Y of the scudy variate y based on Simple Random Sampling without replacement in each stratum. 10 (Contd.) on e 1(b) (c) (d) Consider a population U = {u,, Uys eee UN}. Observations of study variatle (y) and auxiliary variable (x) are yu) =p x(u,) =x, 15 1,2, 0,N. N Let P; -+ where x=) X; .A sample of size i=l nis drawn with PPSWR. Show: that for population N total T => Yj the estimator ¥ 1 an unbiased estimator. Obtain the variance of the estimator of the population to-al. 10 Verify whether the following arrangement is a symmetrica! BIB design. Consider each set a block : (12, 0, 2, 8), (6, 7, 9, 2), (4. 5, 7, 0), (5, 6, 8. 1), (1, 12, 1, 7), 3, 4, 6, 12), (10, 11, 0, 6), (9, 10, 12, 5), (2,3, 5, 11), (1,2, 4, 10), (8, 9, 11. 4), (7, 8, 10, 3), (0, 1, 3, 9). Is it a connected BIBD ? 10 32 plots are arranged in the form ofa4d x8 rectangle. Give a design for 2° factorial experiment with factors A, B, C, D and E in these plots confounding the effects ABC. CDE, ABDE with the rows and AB, CD, ABCD, ExDE, ADE, BCE, ACE with the columns. 10 7 (Contd.)5. SECTION—B Answer any FIVE parts from the following : 5x8=40 (a) (b) (©) @ (e) (f) (a) List the main components of a time series. Explain the method of link relative for measurement of seasonal fluctuations of a time ser:es. 8 Discuss the considerations in the choice of (i) base period and (ii) the formula (method), for constructing price index numbers. 8 Describe exact and near multicollinearity in a regression model. Discuss, along wi-a justification, the effect of these situations on (i) sampling variance and (ii) prediction. 8 Describe the problem of heteroscedastcity in linear regression models. Outline any ome method for overcoming this problem. 8 Explain the use of Pareto curve in the study of income distribution. 8 Describe in detail Indirect Least Scuares method for estimating structural parameters. [s the estimator unbiased and/or consistent ? 8 State the problem of autocorrelatioa in a general linear model (GEM). Why does this problem arise 2 Demonstrate its effect on ordinary least squares estimator (OLSE). 10 (Contd.)(b) @ (a) What do you mean by forecasting in economic models ? Define best linear un siased predictor and obtain the same in a simple linear regression model satisfying al. the basic assum>tions. 10 Explain variate-difference method for trend analysis. How is the appropriate order of differencing determined ? 10 In the constru:tion of price index numbers, exp.ain the terms ‘time reversal test’ aad ‘factor reversal test’. Show that both these tests are satisfied by the Fisher incex. 10 Explain the identification probl2m in a system of simultaneous equations. State, without proof, the rank and order conditions for identifiability of an equation. Examine for identifiability of the following model : Demand : M, =By +B, Y, ~B,R, +B;P,+U,, Supply : M, = Oy +a, Y, +U,, where M = money, Y = income, R = rate of interest. P = p-ice. Assume that R and P are predetermined. 10 9 (Contd ) :(b) ©) (d) Describe the two-stage least squares (2SLS) procedure for structural estimation in < simultaneous equations model. Show that it coincides with indirect least squzres method wher. the equation is exactly ident.fied. 1¢ Discuss how the Engel curve is constructed on the basis of family budget data. Explain and interpret the Engel law in this context How are variations in household size and composition handled ? 10 Write down the auto correlation function of order K. for an AR(1) model X, = ).7X,_, {,} is a white noise process. Shcw that this model +é€, where can be expressed as a moving average process of infinite order. Check the model for stationarity. 10 The following information concerns changes in price and consumption (quantity” of certain major components of the consumption-basket of the labour class : SI. No. Year 2000 Year 2014 Item |Unit | Price] Consumption] Price] Consumption Rs.) cRs.) Rice | Quintal | 500 16 640 20 Wheat | Quintal | 240 12 320 10 Cloth | Metre 16 50 | 20 35,(b) (4) Compute price index using : (i) Fisher’s method; (ii) Marshall-Edgeworth method. Also interpret the results. 1) Discuss the main aspects of tae sampling desiga used by the NSSO for collection of data on household consumer expenciture. If you were to improve upon this design, explain wtat modifications you would suggest, with a ckar justification. lo Define periodogram. State the coanection between the periodogram and the autocovariance funct.on of a time series. Show that h(w’ defined by h(w)= (2) > K(t) exp (-iwt) ONE tame is of period 2x, where K(.) is the auto covariarce function. 10 Define partial auto correlaticn function (PACF). Determine this function for the model : X,=0.7%,, +03, + €, where {¢,} is a sequence of uncorrelated randcm variables with mean zero anc vaziance o - 19